U.S. patent application number 15/511258 was filed with the patent office on 2017-08-31 for identification of a pantograph represented in an image.
The applicant listed for this patent is DTI Group Limited. Invention is credited to Brett Adams, William Hock Oon Lau, En Peng.
Application Number | 20170249750 15/511258 |
Document ID | / |
Family ID | 55532336 |
Filed Date | 2017-08-31 |
United States Patent
Application |
20170249750 |
Kind Code |
A1 |
Peng; En ; et al. |
August 31, 2017 |
IDENTIFICATION OF A PANTOGRAPH REPRESENTED IN AN IMAGE
Abstract
Pantograph identification methods and devices including
computer-implemented methods, software, computer systems for
identifying a pantograph of an electric vehicle represented in an
image captured by a camera. The method includes, for each pair of
adjacent edges represented in the image, determining distances
between the adjacent edges, wherein the distances are in a same
direction for each of the distances. Then determining a point
weight for points of the image associated with the distance by
comparing the distance to a value or a value range representing a
dimension of the pantograph. Further determining a region of the
image that represents the pantograph based on the point
weights.
Inventors: |
Peng; En; (Western
Australia, AU) ; Lau; William Hock Oon; (Western
Australia, AU) ; Adams; Brett; (Western Australia,
AU) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
DTI Group Limited |
Perth Airport, Western Australia |
|
AU |
|
|
Family ID: |
55532336 |
Appl. No.: |
15/511258 |
Filed: |
September 15, 2015 |
PCT Filed: |
September 15, 2015 |
PCT NO: |
PCT/AU2015/050545 |
371 Date: |
March 15, 2017 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
B60M 1/28 20130101; G06K
2209/19 20130101; G06T 2207/30136 20130101; G06T 7/13 20170101;
G06T 7/0006 20130101; B60L 3/12 20130101; G06K 9/00664 20130101;
G06K 9/4604 20130101; H04N 7/183 20130101; G06T 2207/30252
20130101; G06T 7/70 20170101; B60L 2200/26 20130101 |
International
Class: |
G06T 7/70 20060101
G06T007/70; H04N 7/18 20060101 H04N007/18; G06T 7/13 20060101
G06T007/13; B60L 3/12 20060101 B60L003/12; B60M 1/28 20060101
B60M001/28 |
Foreign Application Data
Date |
Code |
Application Number |
Sep 15, 2014 |
AU |
2014903664 |
Claims
1. A computer-implemented method for identifying a pantograph
represented in an image comprised of points, the method comprising:
for each pair of adjacent edges represented in the image,
determining a plurality of distances between the adjacent edges,
wherein the plurality of distances are in a same direction; for
each of the plurality of distances, determining a point weight for
points of the image associated with the distance by comparing the
distance to a value or a value range representing a dimension of
the pantograph; and determining a region of the image that
represents the pantograph based on the point weights.
2. The computer-implemented method according to claim 1, further
comprising: storing an indication in a memory to indicate the
region of the image.
3. The computer-implemented method according to claim 1, wherein
determining the point weight for the points of the image associated
with the distance comprises: determining a positive point weight
for the points associated with the distance if the distance is
within the value range; and determining a negative point weight for
the points associated with the distance if the distance is outside
the value range.
4. The computer-implemented method according to claim 1, wherein
the points associated with the distance are the points located on
or near a line segment that forms the distance between the pair of
adjacent edges.
5. The computer-implemented method according to claim 1, wherein
determining the region of the image based on the point weights
comprises: determining a sum of point weights of points in the
region of the image is greater than a sum of point weights of
points in other regions of the image.
6. The computer-implemented method according to claim 1, wherein
the same direction is substantially perpendicular to an edge
representative of a top of the pantograph represented in the
image.
7. The computer-implemented method according to claim 1, further
comprising: determining a quantity or proportion of points of the
image that have an associated brightness below a first threshold;
and if the quantity or proportion of points of the image is greater
than a second threshold, aborting the method.
8. The computer-implemented method according to claim 1, further
comprising: determining a quantity of edges in the image; and if
the quantity of the edges in the image is greater than a third
threshold, aborting the method.
9. The computer-implemented method according to claim 1, wherein
the points that the image is comprised of comprise one or more
pixels.
10. The computer-implemented method according to claim 1, wherein
the adjacent edges do not have any edge therebetween.
11. The computer-implemented method according to claim 1, wherein
the region tightly contains the pantograph represented in the
image.
12. A computer software program, including machine-readable
instructions, when executed by a processor, causes the processor to
perform the method of claim 1.
13. A computer system for identifying a pantograph represented in
an image comprised of points, the computer system comprising: a
memory to store instructions; a bus to communicate the instructions
from the memory; a processor to perform the instructions from the
memory communicated via the bus for each pair of adjacent edges
represented in the images, to determine a plurality of distances
between the adjacent edges, where the plurality of distances are in
a same direction; for each of the plurality of distances, to
determine a point weight for points of the image between the
adjacent edges in the same direction by comparing the distance to a
value or a value range representing a dimension of the pantograph;
and to determine a region of the image that represents the
pantograph based on the point weights.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] The present application claims priority from Australian
Provisional Patent Application No 2014903664 filed on 15 Sep. 2014,
the content of which is incorporated herein by reference.
TECHNICAL FIELD
[0002] The present disclosure generally relates to pantograph
identification methods and devices. The present disclosure includes
computer-implemented methods, software, computer systems for
identifying a pantograph represented in an image.
BACKGROUND
[0003] An electric vehicle, for example, an electric train, may
have a pantograph installed on the top of the vehicle to contact a
power supply line positioned over the electric vehicle. The
pantograph introduces electric power from the power supply line to
drive the electric vehicle. Sparks may occur around the contact
between the pantograph and the power supply line when the vehicle
is traveling. These sparks may be due to the contact between the
pantograph and the power supply line not being smooth.
[0004] The unsmooth contact indicates either the pantograph or the
power supply line may have been damaged over time. It is
undesirable to operate the electric vehicle with the damaged
pantograph or the power supply line as accidents involving the
electric vehicle may occur. Therefore, the pantograph may be
monitored by a camera capturing images of the pantograph when the
electric vehicle is in operation.
[0005] Throughout this specification the word "comprise", or
variations such as "comprises" or "comprising", will be understood
to imply the inclusion of a stated element, integer or step, or
group of elements, integers or steps, but not the exclusion of any
other element, integer or step, or group of elements, integers or
steps.
[0006] Any discussion of documents, acts, materials, devices,
articles or the like which has been included in the present
disclosure is not to be taken as an admission that any or all of
these matters form part of the prior art base or were common
general knowledge in the field relevant to the present disclosure
as it existed before the priority date of each claim of this
application.
SUMMARY
[0007] There is provided a computer-implemented method for
identifying a pantograph represented in an image comprised of
points, the method comprising:
[0008] for each pair of adjacent edges represented in the image,
determining a plurality of distances between the adjacent edges,
wherein the plurality of distances are in a same direction;
[0009] for each of the plurality of distances, determining a point
weight for points of the image associated with the distance by
comparing the distance to a value or a value range representing a
dimension of the pantograph; and determining a region of the image
that represents the pantograph based on the point weights.
[0010] It is an advantage of the invention that the pantograph
represented in the image may be identified fast and accurately.
[0011] The computer-implemented method may further comprise storing
an indication in a memory to indicate the region of the image.
[0012] Determining the point weight for the points of the image
associated with the distance may comprise determining a positive
point weight for the points associated with the distance if the
distance is within the value range, and determining a negative
point weight for the points associated with the distance if the
distance is outside the value range.
[0013] The points associated with the distance may be the points
located on or near a line segment that forms the distance between
the pair of adjacent edges.
[0014] Determining the region of the image based on the point
weights may comprise determining a sum of point weights of points
in the region of the image is greater than a sum of point weights
of points in other regions of the image.
[0015] The same direction may be substantially perpendicular to an
edge representative of a top of the pantograph represented in the
image.
[0016] The computer-implemented method may further comprise
determining a quantity or proportion of points of the image that
have an associated brightness below a first threshold; and if the
quantity or proportion of points of the image is greater than a
second threshold, aborting the method.
[0017] The computer-implemented method may further comprise
determining a quantity of edges in the image; and if the quantity
of the edges in the image is greater than a third threshold,
aborting the method.
[0018] The points that the image is comprised of may comprise one
or more pixels.
[0019] The adjacent edges do not have any edge therebetween.
[0020] The region may tightly contain the pantograph represented in
the image. For example, the region may be sized to fit
substantially the pantograph represented in the image and as few
other features of the image as possible.
[0021] There is provided a computer software program, including
machine-readable instructions, when executed by a processor, causes
the processor to perform one or more methods described above.
[0022] There is provided a computer system for identifying a
pantograph represented in an image comprised of points, the
computer system comprising:
[0023] a memory to store instructions;
[0024] a bus to communicate the instructions from the memory;
[0025] a processor to perform the instructions from the memory
communicated via the bus
[0026] for each pair of adjacent edges represented in the images,
to determine a plurality of distances between the adjacent edges,
where the plurality of distances are in a same direction;
[0027] for each of the plurality of distances, to determine a point
weight for points of the image between the adjacent edges in the
same direction by comparing the distance to a value or a value
range representing a dimension of the pantograph; and
[0028] to determine a region of the image that represents the
pantograph based on the point weights.
BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Features of the present disclosure are illustrated by way of
non-limiting examples, and like numerals indicate like elements, in
which:
[0030] FIG. 1 is an diagram of an example vehicle system according
to the present disclosure;
[0031] FIG. 2 shows an example method for identifying a pantograph
represented in an image according to the present disclosure;
[0032] FIG. 3 shows an example method for identifying a pantograph
represented in an image according to the present disclosure;
[0033] FIG. 4(a) is an example image of a pantograph that is
captured by a camera;
[0034] FIG. 4(b) is an example image representative of edges in
FIG. 4(a);
[0035] FIG. 4(c) is an example image illustrating points of
interest in FIG. 4(b);
[0036] FIG. 4(d) is an example image with an indication indicative
of an identified pantograph;
[0037] FIG. 5 illustrates an example process for identifying a
pantograph represented in an image according to the present
disclosure;
[0038] FIG. 6(a) shows an example image captured when an electric
vehicle travels in a dark tunnel;
[0039] FIG. 6(b) shows an example image captured when the electric
vehicle travels under a bridge;
[0040] FIG. 6(c) shows an example image illustrating edges
extracted from the example image shown in FIG. 6(b);
[0041] FIG. 7 is an example processing device for identifying a
pantograph represented in an image according to the present
disclosure; and
[0042] FIG. 8(a) and (b) illustrate examples of reducing a search
space for the pantograph represented in an image.
BEST MODES OF THE INVENTION
[0043] FIG. 1 is an diagram of an example vehicle system 100
according to the present disclosure. The vehicle system 100
comprises an electric vehicle 110 and a power supply line 120. The
electric vehicle 110 comprises a pantograph 130. It should be noted
that although only one car of the electric vehicle 110 is shown in
FIG. 1, the electric vehicle 110 may comprise a plurality of
cars.
[0044] The power supply line 120 is an overhead power line that is
installed over the travel path of the electric vehicle 110. The
power supply line 120 comprises a contact wire 140 and a catenary
150.
[0045] The contact wire 140 carries electric power and contacts the
pantograph 130 of the electric vehicle 110, particularly, the
carbon strip part at the top surface of the pantograph 130. The
electric power carried on the contact wire 140 is introduced or
collected to a driving mechanism of the electric vehicle 110, for
example, an electric motor (not shown in FIG. 1), through the
pantograph 130 to drive the electric vehicle 110 on rails 160. In
other examples, the electric vehicle 110 may travel without use of
the rails 160.
[0046] To keep the contact wire 140 within defined geometric
limits, the catenary 150 is used to support the contact wire 140
from above through dropper wires 170. That is, the contact wire 140
is positioned lower than the catenary 150 in this example.
[0047] In FIG. 1, the dropper wires 170 vertically extend between
the contact wire 140 and the catenary 150. The dropper wires 170
attach the contact wire 140 and the catenary 150 at specified
intervals.
[0048] The power supply line 120 is hung over the electric vehicle
110 by suspension cables 180, which may be in turn secured to
support mechanisms (not shown in FIG. 1), for example support
towers or support poles, which are installed along the travel
path.
[0049] In the example shown in FIG. 1, the suspension cables 180
are attached to the power supply line 120 at suspension points
181.
[0050] A camera 190 is installed on the top of the electric vehicle
110 to monitor the pantograph 130. Specifically, the camera 190 may
capture images of the pantograph 130 when the electric vehicle 110
is in operation. The images may be still images and may form part
of a video. An example image 410 of the pantograph 130 captured by
the camera 190 is shown in FIG. 4(a). The image 410 may be formed
by points, which may be one or more pixels of digital images.
[0051] The images of the pantograph 130 captured by the camera 190
are sent to a processing device 191 for further analysis to
determine operation conditions of the electric vehicle 110.
[0052] It should be noted that although the processing device 191
in FIG. 1 is located in the electric vehicle 110, the processing
device 191 may also be located remotely from the electric vehicle
110, or both and each processing device 191 can perform part of the
method of identifying the pantograph 130. Further, although the
pantograph 130 and the camera 190 monitoring the pantograph 130 are
installed on the same car of the electric vehicle 110, as shown in
FIG. 1, the pantograph 130 and the camera 190 may also be installed
on different cars of the electric vehicle 110.
[0053] A method for identifying the pantograph 130 represented in
the image 410 is described with reference to FIGS. 2 and 3.
[0054] Upon receipt of the image 410 from the camera 190,
optionally, the processing device 191 extracts edges from the image
410. Specifically, the processing device 191 may use an edge
detector for example a Canny edge detector, described in John
Canny, "A Computational Approach to Edge Detection," IEEE
Transactions on Pattern Analysis and Machine Intelligence, vol. 8,
no. 6, pp. 679-698, June 1986, doi:10.1109/TPAMI.1986.4767851, to
extract the edges from the image 410. The edges extracted by the
edge detector may have single point or pixel width. As a result, an
image 420 is generated that is representative of the edges in the
image 410, as shown in FIG. 4(b). Alternatively, the image received
may already include a representation of the edges in the image and
the above edge extracting step is not performed accordingly.
[0055] It can be seen from FIG. 4(b) that two edges in the edges in
the image 420 may be adjacent edges. Specifically, "adjacent edges"
in the present disclosure refer to a pair of edges without any edge
therebetween when viewed along a line. Take four edges 1, 2, 3 and
4 shown in FIG. 4(b) as an example, the edges 1 and 2, the edges 2
and 3, the edges 3 and 4 are three pairs of adjacent edges along a
line 4201, shown as a dashed line in FIG. 4(b). The line may be a
vertical straight line with one-point width. It should be noted
that a pair of edges may be adjacent edges when viewed along a
line, but the pair of edges may not be adjacent edges when viewed
along another line.
[0056] An important characteristic of the pantograph 130 in this
example is that the pantograph 130 has a horizontal length and a
dimension, for example thickness. The dimension substantially does
not change along the horizontal length of the pantograph 130. The
dimension of the pantograph 130 may be represented by a value range
[D.sub.min, D.sub.max].
[0057] The dimension may not necessarily be the actual physical
size of the pantograph 130. For example, the dimension may be the
size of the pantograph 130 in the image 420, which may be measured
by points or pixels which the image 420 is comprised of.
[0058] The value range may be pre-determined empirically. The
dimension of the pantograph 130 may also be represented by a value
based on which the value range is derived. For example, the upper
limit of the value range may be 120% of the value, while the lower
limit of the value range may be 70% of the value.
[0059] The processing device 191 scans the image 420 in the
direction of the line 4201 and along the line 4201. In the example
shown in FIG. 4(b), the line 4201 is a vertical line that is
substantially perpendicular to an edge representative of the top of
the pantograph 130 represented in the image 420. In this example,
the line 4201 is a one-point width line.
[0060] The processing device 191 identifies a plurality of pairs of
adjacent edges along the line 4201 in the image 420, for example,
adjacent edges 1 and 2, 2 and 3, 3 and 4, and determines a distance
in the direction of the line 4201 between each of the adjacent edge
pairs along the line 4201. For example, the distance between the
edges 2 and 3 along the line 4201 is D.sub.1.
[0061] The processing device 191 repeats the above process but
along a different line 4202, which is in the same direction as the
line 4201 but is slightly offset from the line 4201. For example,
the line 4202 is offset from the line 4201 by the width of the line
4201, which is one point in this example.
[0062] As a result, the processing device 191 can determine a
distance in the direction of the line 4202, which is the same as
the line 4201, between each of the adjacent edge pairs along the
line 4202. For example, the distance between the adjacent edges 2
and 3 along the line 4202 is D.sub.1'.
[0063] In this example, the processing device 191 scans the image
420 along all vertical lines parallel with the lines 4201 and 4202
as described above. These vertical lines may be evenly spaced or
may be spaced apart in a way that is not even. This way, for each
pair of adjacent edges represented in the image, the processing
device 191 can determine 210 a plurality of distances between the
adjacent edges in the same direction. In other examples, the
scanning process described above may only be performed on a portion
of the image 420 in which the pantograph 130 is anticipated to be
located in to reduce the computing burden of the processing device
191 and speed up the scanning process.
[0064] For each of the plurality of distances between the adjacent
edges, the processing device 191 determines 220 a point weight for
points of the image associated with the distance by comparing the
distance to the value or the value range representing the dimension
of the pantograph. In the present disclosure, the points associated
with the distance are the points located on or near a line segment
that forms the distance between the pair of the adjacent edges.
[0065] Take the distances between the adjacent edges 2 and 3 as an
example, the processing device 191 compares each of the distances,
for example D.sub.1 and D.sub.1', with the value range [D.sub.min,
D.sub.max] that represents the thickness of the pantograph 130.
[0066] The points associated with the distance D.sub.1 are the
points located on or near a line segment of the line 4201. One end
point of the line segment is the intersection point between the
line 4201 and the edge 2, and another end point of the line segment
is the intersection point between the line 4201 and the edge 3. The
length of the line segment forms the distance between the pair of
the adjacent edges 2 and 3 along the line 4201.
[0067] Similarly, the points associated with the distance D.sub.1'
are the points located on or near a line segment of the line 4202.
One end point of the line segment is the intersection point between
the line 4202 and the edge 2, and another end point of the line
segment is the intersection point between the line 4202 and the
edge 3. The length of the line segment forms the distance between
the pair of the adjacent edges 2 and 3 along the line 4202.
[0068] If the distance is within the value range [D.sub.min,
D.sub.max], the points associated with the distance are identified
as points of interest and a positive point weight may be determined
310 for these points. On the other hand, a negative point weight
may be determined 320 for these points if the distance is outside
the value range [D.sub.min, D.sub.max]. In this example, the
positive point weight for the points of interest is +10, and the
negative point weight for non-points of interest is -1.
[0069] Therefore, the point weight for a point (p) may be expressed
by the following equation (1):
weight ( p ) = { + 10 , p is a point of interest - 1 , otherwise (
1 ) ##EQU00001##
[0070] In the above three pairs of adjacent edges, the distance
between the adjacent edges 3 and 4 along the line 4210 is within
the value range [D.sub.min, D.sub.max], while the distance between
the adjacent edges 1 and 2 or 2 and 3 is outside the value range
[D.sub.min, D.sub.max]. As a result, the points associated with the
distance between the adjacent 3 and 4 along the line 4201 are
identified as the points of interest and the positive point weight
of +10 is assigned to the points, while the negative point weight
of -1 is assigned to the points between the adjacent edges 1 and 2
and the adjacent edges 2 and 3.
[0071] The processing device 191 may identify all the points of
interest and determine the point weighs for all the points in the
image 420.
[0072] For illustration purposes, an image 430 illustrates the
points of interest in the image 420, in which the grey portion
represents the points of interest having the point weight of +10,
while the black portion represents the non-points of interest
having the point weight of -1, as shown in FIG. 4(c). As can be
seen from FIG. 4(c), some of the points of interest belong to
points that constitute the pantograph 130. At the same time, some
of the points of interest do not belong to the pantograph 130, for
example, the points of interests that are located at the
bottom-left corner of the image 430.
[0073] The processing device 191 then determines 230, based on the
point weights for the points in the image 420, a region
R.sub.optimal of the image 420 that represents the pantograph 130.
Specifically, the processing device 191 determines 330 a sum of
point weights of points in the region of the image 420 is greater
than a sum of point weights of points in other regions of the image
420. The sum of point weights of points in the region R.sub.optimal
that is acceptable is in an expected range. Preferably, the region
is a region that tightly contain the pantograph 130.
[0074] To determine the region R.sub.optimal, a quality function
f(R) of an image region R in an image I is defined by the following
equation (2):
f ( R ) = p .di-elect cons. R weight ( p ) ( 2 ) ##EQU00002##
[0075] The result of quality function f(R) represents a sum of
point weights of the points in the image region R.
[0076] The region R.sub.optimal may be defined by the following
equation (3):
R optimal = R I argmax f ( R ) ( 3 ) ##EQU00003##
[0077] That is, the region R.sub.optimal represents a region in the
image I that has a maximum sum of point weights. The region
R.sub.optimal may be obtained by applying a sub-window search
algorithm for example I-ESS described in An, Senjian An, P.
Peursum, Wanquan Liu, S. Venkatesh, "Efficient algorithms for
subwindow search in object detection and localization" cvpr, pp.
264-271, 2009 IEEE Conference on Computer Vision and Pattern
Recognition, 2009.
[0078] By applying the equations (2) and (3) to the point weights
of points, as illustrated by the image 430, the region
R.sub.optimal may be obtained, as shown in an image 440 of FIG.
4(d). As can be seen from the image 440, the region R.sub.optimal
contains the pantograph 130 and the boundary of the region
R.sub.optimal, represented by a line box, tightly surrounds the
pantograph 130. Therefore, the pantograph 130 is identified.
[0079] The processing device 191 then stores an indication to
indicate the points in the region R.sub.optimal. As the region
R.sub.optimal is bounded by the line box having four corners, as
shown in the image 440, positions of the four corners may be used
as the indication to indicate the points in the region
R.sub.optimal. The points in the region R.sub.optimal may also be
indicated in other ways without departing from the scope of the
present disclosure.
[0080] FIG. 5 illustrates an example process 500 for identifying a
pantograph represented in an image, which may be performed by the
processing device 191 shown in FIG. 1.
[0081] In practice, the electric vehicle 110 may travel at night or
in a dark tunnel, it is thus difficult to extract edges from the
images captured by the camera 190 due to the low contrast with the
dark background.
[0082] FIG. 6(a) shows an example image 610 that is captured by the
camera 190 when the electric vehicle 110 travels in a dark
tunnel.
[0083] It can be seen from the image 610 that the pantograph in the
image 610 has very low contrast with the dark background, which
makes the pantograph difficult to be identified.
[0084] The electric vehicle may also travel under a bridge or pass
by a depot. In such a case, the background of the image captured
may contain excessive man-made patterns that may appear to be the
power supply line 120 or the pantograph 130 when edges are
extracted. These patterns are likely to cause false identification
of the pantograph 130.
[0085] FIG. 6(b) shows an example image 620 that is captured by the
camera 190 when the electric vehicle 110 travels under a bridge.
FIG. 6(c) shows an example image 630 that illustrates the edges
extracted from the image 620.
[0086] It can be seen from the image 630 that the edges of the
man-made patterns in the image 620, for example, windows and poles
of the bridge, make it difficult to distinguish the pantograph 130
from the image 630.
[0087] Therefore, in the process 500, as the electric vehicle 110
travels, the processing device 191 obtains Global Positioning
System (GPS) data 501 indicative of the current geographic location
of the electric vehicle 110. Meanwhile, the processing device 191
checks a database that includes locations of bridges and depots to
determine 505 if the electric vehicle 110 is under a bridge or in a
tunnel or a depot. In another example, the GPS data and the
database including locations of bridges and depots may not be
needed without departing from the scope of the present
disclosure.
[0088] If the current geographic location of the electric vehicle
110 indicates that the electric vehicle 110 is under a bridge or in
a tunnel or a depot, the processing device 191 aborts 527 the
process 500 or discard the image that has already been
captured.
[0089] If it is indicated from the GPS data 501 that the electric
vehicle 110 is not under a bridge or in a tunnel or a depot, the
processing device 110 may instruct the camera 190 to capture 507 an
image or proceed to processing the image that has already been
captured by the camera 190.
[0090] Additionally or alternatively, the processing device 191
converts 509 the captured image to a grey level image. Before
extracting edges from the grey level image, the processing device
191 determines if there is a sufficient contrast 511 between the
pantograph and the background in the grey level image.
[0091] To do this, the processing device 191 determines a quantity
or proportion of points of the image that have an associated
brightness below a first threshold. If the quantity or proportion
of the points of the image is greater than a second threshold,
indicating that the image may be too dark, as shown in FIG. 6(a),
the processing device 191 aborts 527 the method.
[0092] Specifically, the processing device 191 checks the
brightness level, for example, the grey level, of a point in the
grey level image, if the brightness level of the point is below 20%
of the maximum brightness level, indicating the point is a dark
point, an counter is increased by one. The processing device 191
may repeat the above procedure for every point in the grey level
image. If the resulting counter is greater than 30% of the number
of the points in the grey level image, indicating more than 30%
percent of the points in the grey level image are dark points, the
contrast of the grey level image is not acceptable for further
processing, the processing device 191 aborts 527 the process
500.
[0093] To improve the processing speed, the processing device 191
may only determine the contrast for part of the grey level image
that contains the pantograph, for example, the top half of the grey
level image.
[0094] In another example, the processing device 191 may calculate
the standard deviation of the brightness levels of the points in
the grey level image to determine if the contrast of the grey level
image is acceptable. Particularly, if the standard deviation is
below a threshold, the contrast of the image may not be considered
to be acceptable. This way, only one threshold is needed.
[0095] If the contrast of the grey level image is acceptable, the
processing device 191 extracts 513 edges from the grey level image
to generate an image (for example, the image 420 shown in FIG.
4(b)) representative of the edges in the grey level image and
determines 515 if a quantity of edges in the image is acceptable.
If the quantity of the edges in the image is greater than a third
threshold, indicating the image may contain excessive edges that
may cause false identification of the pantograph, the processing
device 191 aborts 527 the process 500.
[0096] The quantity of the edges in the image may be represented by
the number of points that belong to edges in the image. The third
threshold may represent the proportion of edge points in the image.
If the proportion of edge points is more than 10% of the total
number of points in the image, it is considered that the quantity
of edges in the image is not acceptable.
[0097] To improve the processing speed, the processing device 191
may only determine the quantity of the edges for part of the image
that contains the pantograph, for example, the top half of the
image.
[0098] If the quantity of edges in the image is acceptable, the
processing device identifies 517 points of interest, determines 519
point weights for the points of interest and non-points of
interest, and searches 521 for the region R.sub.optimal that
contains the pantograph 130 according to the methods described with
reference to FIGS. 2 and 3.
[0099] If the result of the quality function for the region
R.sub.optimal, i.e., the sum of point weights of the points in the
region R.sub.optimal, is in an expected range, the region
R.sub.optimal is acceptable 523 as a region that contains the
pantograph 130.
[0100] The pantograph 130 is then identified 525 by using for
example the line box representing the region R.sub.optimal, as
shown in the image 440 of FIG. 4(d). Otherwise, the processing
device 191 aborts 527 the process 500.
[0101] To further improve the accuracy of the methods and processes
described above, one or more of the following processes may be
performed.
Determination of the Value Range Representing the Dimension of the
Pantograph
[0102] There are two parameters used in the above methods:
D.sub.min and D.sub.max. To determine the values of these
parameters, two or four reference images are employed depending on
the degree of freedom of the pantograph 130 relative to the camera
190 monitoring the pantograph 130.
[0103] When the pantograph 130 and the camera 190 are installed on
the same car of the electric vehicle 110, the pantograph 130 may
substantially only move up and down in the images captured by the
camera 190. In this case, two reference images are needed to
determine D.sub.min and D.sub.max with one reference image showing
that the pantograph 130 is located at the bottom of the image and
the other one showing the pantograph 130 is located at the top of
the image.
[0104] When the pantograph 130 and the camera 190 are installed on
different cars of the electric vehicle 110, the pantograph 130 can
move in any direction especially when the electric vehicle 110 is
turning. In this case, four reference images are needed to
determine D.sub.min and D.sub.max. Each of the four reference
images shows the pantograph is located at one of extreme locations,
for example, top, bottom, four corners of the images: top left, top
right, bottom left and bottom right.
[0105] For both scenarios, the pantograph 130, particularly, the
carbon strip part, is marked by using a bounding box in each
reference image. In each bounding box, an edge detection process is
performed to extract edges in the bounding box. Along each column
of points in the bounding box, the points are scanned to identified
adjacent edges. For each pair of adjacent edges, a distance between
the adjacent edges along the column is determined and a vote is
placed on a distance bin. In the present disclosure, a distance bin
refers to a certain distance range. For example, a distance bin i
may refer to a distance range between 2i and 2 (i+1). If a distance
is 9, the distance may result in a vote into the distance bin No. 4
since the distance is in the distance range between 2.times.4=8 and
2.times.(4+1)=10. As a result, neighbouring distance bins may
receive a majority of the votes. Therefore, the value range
representing the dimension of the pantograph 130 may be determined
according to the distance ranges corresponding to these distance
bins.
Fixed Size Sub-Window
[0106] In searching for the region R.sub.optimal that contains the
pantograph 130, a fixed size sub-window may be used in the present
disclosure for the sub-window search algorithm to reduce the
computing burden of the processing device 191.
[0107] For each scenario, since the bounding boxes for the
reference images may represent the extreme sizes, i.e., the maximum
size and the minimum size, of the pantograph 130, the size of the
largest bounding box may be used as the size of the fixed size
sub-window.
[0108] As a result, the sub-window search algorithm may be
performed with the fixed size sub-window, and can be completed with
less computing capabilities and within less time.
Reduction of the Search Space
[0109] In searching for the region R.sub.optimal that contains the
pantograph 130, the search space for the sub-window search
algorithm may be the entire image 420, shown as an image area 801
in FIGS. 8(a) and (b).
[0110] In the present disclosure, the search space may be reduced
based on the location of the pantograph 130 in the image area 801.
As described above, the location of the pantograph 130 may include
extreme locations in the image area 801, for example, top, bottom,
four corners of the image area 801: top left, top right, bottom
left and bottom right.
[0111] As shown in FIG. 8(a), in the scenario where the pantograph
130 and the camera 190 are installed on the same car of the
electric vehicle 110, as described above, since the pantograph 130
may substantially only move vertically in relation to the camera
190, the two reference images show that the pantograph 130 is
located at the top and bottom of the image area 801, represented by
a top bounding box 803 and a bottom bounding box 805 in the image
area 801. These bounding boxes 803, 805 are indicated by the solid
line boxes in FIG. 8(a).
[0112] If a fixed size sub-window is used, the search space may be
reduced to a line segment between the locations of the centres of
the bounding boxes 803, 805, shown as a search line 807 in FIG.
8(a), and the search may be performed on the search line.
[0113] On the other hand, if the size of sub-window is not fixed,
the search space may be reduced to an area defined the extreme
locations of the pantograph 130. In this example, the reduced
search space may be a search area 809 tightly enclosing the extreme
locations of the pantograph 130, indicated by the dash line box in
FIG. 8(a).
[0114] As shown in FIG. 8(b), in the scenario where the pantograph
130 and the camera 190 are installed on different cars of the
electric vehicle 110, the four reference images show that the
pantograph 130 is located at the top left, top right, bottom left
and bottom right of the image area 801, represented by a top-left
bounding box 802, a top-right bounding box 804, a bottom-left
bounding box 806 and a bottom-right bounding box 808.
[0115] If a fixed size sub-window is used, the search space may be
reduced to a search area 810 that is bounded by the centre
locations of the four bounding boxes 802, 804, 806, 808, indicated
by the inner dash line box in FIG. 8(b).
[0116] On the other hand, if the size of sub-window is not fixed,
the search space may be reduced to a search area 812 tightly
enclosing the extreme locations of the pantograph 130, indicated by
the outer dash line box in FIG. 8(b).
[0117] This way, the processing device 191 may search the reduced
search space for the region R.sub.optimal. As a result, the
sub-window search algorithm may be performed within the reduced
search space, and can be completed with less computing capabilities
and within less time.
Acceptance of the Region R.sub.optimal
[0118] As described with reference to the step 525 of the process
500 shown in FIG. 5, the region R.sub.optimal is acceptable as the
region that contains the pantograph 130 if the result of the
quality function for the region R.sub.optimal is in the expected
range.
[0119] For each scenario, since the bounding boxes for the
reference images may represent the extreme conditions of the region
R.sub.optimal, the results of the quality functions for the
bounding boxes may be used as the basis of the expected range that
the sum of point weights of the points in the region R.sub.optimal
is in.
[0120] For example, the upper limit of the expected range may be
the maximum value in the results of the quality functions for the
bounding boxes, while the lower limit of the expected range may be
the minimum value in the results. In practice, the upper/lower
limit of the expected range may be relaxed for robustness without
departing from the scope of the present disclosure.
[0121] FIG. 7 illustrates an example processing device 191
according to present disclosure.
[0122] The processing device 191 includes a processor 710, a memory
720 and an interface device 740 that communicate with each other
via a bus 730. The memory 720 stores instructions and data for the
methods and processes described above, and the processor 710
performs the instructions from the memory 720 to implement the
methods and processes. It should be noted that although the
processing device 191 is shown as an independent entity in FIG. 1,
the processing device 191 may also be part of another entity for
example the camera 190.
[0123] The processor 710 may perform the instructions from the
memory 720 communicated via the bus 730 [0124] for each pair of
adjacent edges represented in the images, to determine a plurality
of distances between the adjacent edges, where the plurality of
distances are in a same direction; [0125] for each of the plurality
of distances, to determine a point weight for points of the image
between the adjacent edges in the same direction by comparing the
distance to a value or a value range representing a dimension of
the pantograph; and [0126] to determine a region of the image that
represents the pantograph based on the point weights.
[0127] The processor 710 may also perform other methods and
processes described above with reference to the accompanying
drawings.
[0128] It should be understood that the techniques of the present
disclosure might be implemented using a variety of technologies.
For example, the methods described herein may be implemented by a
series of computer executable instructions residing on a suitable
computer readable medium. Suitable computer readable media may
include volatile (e.g. RAM) and/or non-volatile (e.g. ROM, disk)
memory, carrier waves and transmission media. Example carrier waves
may take the form of electrical, electromagnetic or optical signals
conveying digital data streams along a local network or a
publically accessible network such as the Internet.
[0129] It should also be understood that, unless specifically
stated otherwise as apparent from the following discussion, it is
appreciated that throughout the description, discussions utilizing
terms such as "obtaining" or "determining" or "sending" or
"receiving" or the like, refer to the action and processes of a
computer system, or similar electronic computing device, that
processes and transforms data represented as physical (electronic)
quantities within the computer system's registers and memories into
other data similarly represented as physical quantities within the
computer system memories or registers or other such information
storage, transmission or display devices.
[0130] It will be appreciated by persons skilled in the art that
numerous variations and/or modifications may be made to the
above-described embodiments, without departing from the broad
general scope of the present disclosure. The present embodiments
are, therefore, to be considered in all respects as illustrative
and not restrictive.
* * * * *